The present invention relates to bacteriophage tail proteins and the derivatives and fragments thereof that are capable of binding endotoxins in the absence of bivalent positive ions, especially Ca2+ or Mg2+. Further, the present invention relates to methods for depleting endotoxins from solutions and samples using the bacteriophage tail proteins according to the present invention and to a detection method for endotoxins.
Bacteriophages recognize structures (membrane proteins, sugar molecules etc.) on the surface of theirs host bacteria by corresponding proteins, which bacteriophages have on theirs surface. Some bacteriophages have only one type of recognition protein, e.g. salmonella phage P22, others at least two or more. The recognition proteins may have enzymatic activity, as phage P22 (Seckler, J. Struct. Biol. 1998; 122(1-2):216-222), or may not have enzymatic activity. Enzymatic activity means, that these proteins, e.g. the P22 tail spike protein, are able to hydrolyze the receptor molecule, that they recognize, i.e. at p22 the salmonella O-antigen. The most known bacteriophage having two recognition proteins is for E. coli the specific phage T4. This phage has long and short tail fibers. The long tail fibers conduces the specific recognition of its host E. coli by the membrane protein OmpC or by lipopolysaccharide for E. coli B. While the long tail fibers of the phages T4, T2 and K3 bind specifically to OmpC and lipopolysaccaride of E. coli B (T4), respectively, OmpF (T2; Hantke K., Mol Gen Genet. 1978; 164 (2):131-135) and OmpA (K3; Hancock R E, Reeves P., J Bacteriol. 1975; 121(3):983-993; Riede I, Eschbach M L, Henning U., Mol Gen Genet. 1984; 195(1-2):144-152), the short tail fibers are located at the bottom side of the phage and are not involved in the host specificity, but replaceable between T4, T2 and K3 phages (Riede, Mol Gen Genet. 1987; 206(1):110-115). Only after the binding of at least three long tail fibers, the short tail fibers are folded out of the basis plate and are responsible for irreversible binding of the T4 phage to the E. coli hosts (Leiman et al., Cell Mol Life Sci. 2003; 60(11):2356-2370). These short tail fiber proteins bind, as shown for page T4 (WO2004/001418), in the core region of the lipopolysaccharide and thus are qualified for recognizing and binding endotoxin.
Endotoxins (ET) describe a family of lipopolysaccharides, which form the outer cell membrane of gram-negative bacteria together with proteins and phospholipids. Endotoxins only occur in this bacteria group and play an important role at the organization, stability and barrier function of the outer membrane. Numerous bacteriophages use endotoxin and general lipopolysaccharides, respectively, for specific recognition of theirs host bacteria.
Endotoxins are biomolecules which may be found in practically all aqueous solutions without corresponding precautionary measures. Endotoxins effect on human and animals highly pyrogenically, so they cause fever response and are able to result in a sepsis, a heavy dysfunction of the immune system involving a high mortality rate. Therefore contamination with endotoxin, e.g. at the production of proteins for medical or pharmaceutical use, have to be detected exactly and be removed consequently. Endotoxin presents a problem by genetically produced pharmaceuticals, gentherapeutic agents or substances, which are injected into humans or animals (e.g. veterinary treatment or in animal tests). However, not only for medical, but also for research applications, such as transfection experiments of mammalian cells, an inhibition or decrease, respectively, of the transfection efficiency by endotoxin may be found.
All endotoxin variations consist of a heteropolysaccharide, that is covalent bound to lipid A (Hoist, O., 1999, In: Endotoxin in health and disease (Brade, H. et al; eds.), Marcel Dekker Inc. New York)). Lipid A anchors endotoxin in the outer bacteria membrane. The heteropolysaccharide, consisting of a core oligosaccharide and the O-antigen, points to the ambient solution and determines the serological identity of the bacterium. The O-antigen consists of repetitive oligosaccharide units, whose composition is specific for each strain (see Hoist et al., supra). Characteristic blocks of the core oligosaccharide are 2-keto-3-deoxyoctonic acid (KDO) and L-glycero-D-manno-heptose (Hep).
The most conservative part of different genera of endotoxin is the Lipid A. The inner heart region is related conserved as lipid A, while the outer core region already has a higher variation. The inner heart region, KDO and lipid A carry several phosphate groups as substitutes themselves and are consequently responsible for the negative charge of endotoxin. Furthermore, the phosphate groups of Lipid A and the core region may be substituted with arabinose, ethanolamine and phosphate variably. Single saccharide building blocks of the O-antigen are acetylated, sialylated or glycolysated. The O-antigen varies moreover concerning the amount of repetitive units, wherefore the endotoxin population of each bacterium has a certain heterogeneity (Palva E. T. and Makela P. H., Eur J Biochem. 1980; 107(1):137-43; Goldman R. C. and Leive L., Eur J Biochem. 1980; 107(1):145-53).
To be able to use proteins within clinical studies, the European and American pharmacopoeia demand, that the proteins under-run certain limit values of endotoxin load (e.g. immune serum globulin ≦0.91 EU/ml, this corresponds to ≦5 EU/kg body weight & hour (dose rate=EU/kg*h); EU=endotoxin unit; FDA (Food and Drug Administration): Guideline on Validation of LAL as End Product). In case a drug and therein-containing proteins, respectively, have a too high endotoxin load, it is possible that this induces the death of the patient. The misdirected immune defense damages the patient by an over-reaction. This may induce tissue inflammation, decrease in blood pressure, tachycardia, thrombosis culminating in septic shock and multiple organ failure. Already a long running exposition of endotoxin in picogram quantities may induce chronic side effects e.g. low immunity, septic symptoms etc. Within the substance production, it is tried to deplete and remove, respectively, endotoxin as far as possible, in particular in processes of “Good Manufacturing Practice” (GMP) conditions. However, the removal of endotoxin on proteins, polysaccharides and DNA is problematically. In particular, great problems exist on proteins, because of whose intrinsic properties as charge state or hydrophobicity, which almost inhibit endotoxin removal and may lead to great losses of products, respectively, during the removal process.
Furthermore, the endotoxin detection as well as the removal is affected by the environment, since factors e.g. ion composition, pH-value, temperature or the presence of other substances may influence the interaction of a ligand with endotoxin rigorously. Thereby it must be considered, that the interaction of ligands may be carried out with different structure parts of the endotoxins as the hydrophobic Lipid A or the hydrophilic polysaccharide part. According to this, normally these interactions depend on ionic or hydrophobic forces, which are affected differently by the composition of the solution. The polysaccharide structure of endotoxins is stabilized by bivalent positive ions as calcium or magnesium (Galanos C. and Lüderitz O., Eur. J. Biochem. 1975; 54:603-610). These ions are also able to interfere with ligands (“bridging-effect”).
In general, there are a number of methods for depleting and removing endotoxin, respectively, from biological solutions. However particularly for proteins, there are no general applicable standard methods so far. The used methods are adapted to the special properties of the protein and its corresponding production process in each case. There are different opportunities for depleting endotoxins, wherein each of these methods has specific advantages and disadvantages.
The ultra filtration (Petsch, D. & Anspach, F. B., 2000, J. Biotechnol. 76, 97-119 and references therein) is used for depleting endotoxin from water and solutions with low-molecular substances as salts, sugar and antibiotics. However, it is not qualified for high-molecular proteins or DNA.
The two-phase extraction (e.g. WO 01/66718, Merck) should separate water-soluble proteins and DNA from endotoxin, but it involved detergent residues in the purified product. However, the method is time-consuming because of repeating the purification process for several times.
Likewise, an anion exchanger process (DEAE) (e.g. U.S. Pat. No. 5,990,301, Qiagen; WO 94/14837, Enzon; EP0592989, Braun Melsungen) is used for depleting endotoxins from DNA and acidic proteins, but it requires a low ionic strength (<50 mM NaCl) and leads to a protein co-adsorption of acidic proteins. For alkaline proteins cation-exchanger are used, which partly are combined with detergents (e.g. US 2002/0147315 A1).
Cationic peptides are used for removing endotoxin in EP 0232754 B1 (Commonwealth Biotechnologies).
In addition, hydrophilic matrices are used as a combination of dextran and N′,N′-methylenebisacrylamide (U.S. Pat. No. 5,917,022).
Hydrophobic chromatography methods are used in WO94/14837 (Enzon).
The affinity adsorption (e.g. polymyxin B, histamine, histidine, poly-L-lysine, polyethylenimine) e.g. GB 2,192,633 (Hammersmith Hospital), US2002/0130082 A1 (Tokodoro), U.S. Pat. No. 5,510,242 or WO95/025117 (GMBF) is a further method for depleting endotoxins from DNA and proteins (e.g. BSA, myoglobin, gamma globulin, cytochrome C), but it is toxic in the case of polymyxin B and may lead to a co-adsorption of proteins at low ionic strength.
Following methods describe a removal of endotoxin by means of metal affinity chromatography (U.S. Pat. Nos. 6,365,147; 6,942,802; WO02/083710, American Cyanide).
In addition, LPS-binding proteins or peptides or derivates thereof are used for specific binding of endotoxin (U.S. Pat. No. 6,376,462, Xoma Corp.; U.S. Pat. No. 6,384,188, Dana Faber Cancer Institute; WO95/005393, Morphosys; WO95/008560, Centocor; WO95/025117, Scripps).
Further on, the immune-affinity chromatography is used, wherein the specificity for certain endotoxins can only be achieved by expensive antibodies against core oligosaccharide (U.S. Pat. No. 5,179,018, Centocor; WO 00/08463, Bioserv; EP0074240, Gaffin).
Further, the S3 delta peptide (WO 01/27289) of the factor C (a component of the LAL-test) (WO 99/15676 both: National University of Singapore) is used for proteins (e.g. BSA, chymotrypsinogene), however this method has a low efficiency at high ionic strength and high production costs come along (production of insect cell cultures).
Furthermore the endotoxin neutralizing protein (ENP) from Limulus polyphemus, that also binds specifically to endotoxin (e.g. U.S. Pat. Nos. 5,747,455; 5,627,266) or the LPS binding protein of the horseshoe crab (U.S. Pat. No. 5,760,177) is used for depleting endotoxins. The recovery of this protein from the horseshoe crab or recombinant from saccharomyces is also time-consuming and cost-intensive.
A further method for removing endotoxins from a sample is described in the WO2004/001418. Thereby endotoxins are bound to a carrier immobilized with bacteriophage tail proteins and are so separated from the sample. For an efficient separation, bivalent ions are necessary by what the method cannot be carried out with industrial relevant buffers e.g. phosphate or citrate buffers or in the presence of chelators as EDTA or EGTA.
Essentially three methods exist for protein solutions adapted to the properties of the target proteins in application in pharmaceutical industry:                anion exchange chromatography        reserved-phase chromatography; This has the disadvantage, that it is not suitable for all proteins similarly and for hydrophobic proteins particularly problematically. Furthermore, this method is very time-intensive and proteins are normally denaturated under the conditions of the reserved-phase chromatography, so that they have to be renaturated afterwards time-consuming and often with a high material loss        RemTox (Fa. Milipore): This method has the disadvantage that beside a very long incubation time the unspecific binding fraction is high and the recovery of proteins is often not sufficiently.        
A rough depletion of endotoxin from proteins to a value of up to 10 EU/ml is possible in numerous cases with the existing methods. However, still the remaining concentrations of endotoxin affect toxically. Therefore, a further depletion (i.e. precision purification) is demanded and dependent, respectively, on the protein doses in the medical application. The European pharmacopoeia, the USP (United States Pharmacopeial Convention) and the FDA (Food and Drug Administration) specify the limit values for medical application bindingly (e.g. 5 EU/kg body weight and hour for intravenous applications). However, the precision purification is often not warranted sufficiently with the present methods. The standard methods have relevant disadvantages and are often not applicable for certain proteins or only with a relevant loss of the target protein.
Further, in view of industrial applications it have to be considered, that only buffer substances as phosphate, citrate, borate, carbonate or acetate as cheap as possible are used for reasons of economy. Therefore, the interaction of ligands with endotoxins should not be interfered by these buffers. For binding reactions needing calcium, in particular buffers or additives are problematically, which coordinate calcium as EDTA, EGTA or citrate. In addition, buffers whose salts build insoluble or hardly soluble precipitations with calcium are problematically. For example, calcium phosphate precipitates so there is only a low concentration of free calcium in phosphate buffers.
Beside of depleting and removing endotoxin, respectively, the endotoxin detection in samples, solutions and pharmaceutical preparation plays an important role. Currently six detection methods are described for endotoxin in biologic solutions, wherein only the first two methods are accredited from the FDA. The EAA (endotoxin activity assay) is accredited from the TPD (Therapeutic Product Directorate of Canada) and from the FDA under certain conditions (high risk for sepsis at intensive patients) also. 1. “Rabbit Pyrogen Testing”: A method in which an endotoxin solution is injected to a living rabbit to cause an immune reaction. This immune response caused by endotoxin is verified by fever. 2. Clearly better to standardize is the “Limulus Amoebocyte Lysate (LAL)”-test, which is currently the most applied test (Cambrex-BioWhittacker, Inc., Charles River, Inc., Associates of Cape Cod, Inc., all USA). For this method, an enzyme cascade is induced in the blood of the horseshoe crab (Limulus polyphemus) after the contact of endotoxin. The existence of endotoxin can be measured by four different methods (gel-clot, turbidimetric, calorimetric and chromogenic assay). 3. The InVitro Pyrogene test based on the detection of interleukine-1β in human blood, which is involved in the induction of fever. The test consists of an incubation step of human blood with the examining solution and the following detection of interleukins by antibodies. 4. A similar method is the detection of the induction of prostaglandine (PGE2) in rabbit blood after the contact with endotoxin (Ochiai et al., Microbiol. Immunol., 2003, 47, 585-590). 5. A further possibility is the application of a special cell culture systems (Sterogene Inc., USA) with which the activation of monocytes is pursued by the formation of certain cytokins. 6. The EAA (endotoxin activity assay) by the company Spectral Diagnostics, Inc., Canada is also a blood test. Endotoxin reacts with antibodies, wherein the signal is enforced and detected as chemiluminescence after the complement activation in the patient owned neutrophiles by means of a zymosans.
However, the both first named methods are very expensive and not at least critical for nature conservation reasons because of the high demand of test animals and blood of the very rare horseshoe crab, respectively. In fact, the LAL-test is able to be miniaturized and automated but it has massive disadvantages at the application. It is labor-intensive, requires special trained staff, relative long incubation times, relative big sample volumes and expensive reagents. A onetime opened LAL-solution has to be processed and used up directly, because the components aggregate within a few hours because of low stability. Bivalent ions have to be present in the application of the test, the pH-value is relatively limited (pH 6-7.5) and present glucans often interfere the test. Endotoxin is often masked, i.e. it is e.g. not recognized, if it is bound to proteins. The InVitro Pyrogen test requires as fresh human blood as possible and is relative time-intensive, because the production of the interleukins requires 10 to 24 hours. The main advantage of this method is that also other pyrogens are detected beside endotoxins. This test is primarily intended for replacement of the “Rabbit Pyrogen test”. For all test methods, trained stuff is required and the methods are sensitive for interference, because e.g. it is possible that the immune system of rabbits reacts differently at the same dose of endotoxin. The cell culture method of the company Sterogene is also, as all cell culture methods, very complex and has problems with the standardization. If the different methods for detecting endotoxin are compared, the results often differ from each other, i.e. different endotoxins are not recognized by different test components in the same way. Altogether, it can be fixed, that no easy manageable economic method exists for detecting endotoxin and that the currently used methods have numerous disadvantages.